Will inplace save memory by avoiding x2?

[enfeizhan]

Hope this is the place to ask the developers this question directly. Always wondering which way saves memory. As an example, we use dropna:

df.drop('col', inplace=True, axis=1)

or

df = df.drop('col', axis=1)

[jreback]

no inplace creates a new copy then assigns the pointer

these are equivalent from a memory perspective

[markroxor]

So effectively inplace reduces the extra step and that's the only difference? No gain in terms of resources used? @jreback